Distributive Justice, Dignity and the Lifetime View

Distributive Justice, Dignity and the Lifetime View

Abstract

This paper provides a critical examination of the strongest defences of the pure lifetime view, according to which justice requires taking only people’s whole lives as relevant when assessing and establishing their distributive entitlements and obligations.The paper proposes that we reject a pure lifetime view and replace it with an alternative view, on which some time-specific considerations – that is to say, considerations about how people fare at specific points in time – have non-derivative weight in determining what our obligations are to them.

It seems a fair generalisation to say that very old members of

our society often face greater hardships in life than younger

members. People in their 80s suffer more often from illness, and

many are less wealthy and socially more excluded than people in

their 30s or 40s. Our immediate reaction to this discrepancy

might be to think that there is an injustice here, and that the

young have an obligation to make greater efforts at ameliorating

the conditions of the elderly. But once we reflect on the fact

that the young will themselves be old, and that the old were once

young, we might find ourselves with a different reaction. We

might think that what matters for determining whether the young

have obligations towards the elderly is not how members of

different age groups fare relative to each other in the present, but

whether life as a whole will have been, in some sense, fair to them

all. After all, this case is not dissimilar from another one

where that reaction seems apt: imagine that one group of

travellers is passing through a difficult stage, as another group

are passing through an easy stage, of the same journey. Would

there be anything unfair about that difference between them if

the latter group will soon be passing through the same difficult

stage themselves?

That both of these reactions are reasonable raises a

difficult question for anyone seeking to determine what the young

owe the old, or, more generally, what justice demands between

members of different age groups.1 Those two reactions reflect two

fundamentally different views about how the fact that each

person’s life stretches over time affects what obligations others

have towards her; and any theory of distributive justice, in

order to be complete, needs to take a stance on which of these

2

two views about (what I will call) the temporal subject of justice is

correct.

There are two related respects in which a theory of

distributive justice is incomplete until it provides a view of

the temporal subject of justice. First, as I remarked above, a

theory of distributive justice needs such a view in order to

settle questions about the obligations of justice in a particular

range of cases, namely those involving members of different age

groups. That is, a view on the temporal subject of justice is

needed to widen the range of judgements a theory of justice can

support. But secondly, and more fundamentally, a theory of

distributive justice must presuppose a particular view on the

temporal subject of justice to make any determinate judgement

about what is owed to any individual. For unless we know whether

whole lives or temporal stages of a life are the subject of

justice, general ideals of justice such as, for example, that

people should be equal in their opportunities or outcomes, are

insufficiently determinate: we just don’t know what the ideal of

equality demands until we know whether opportunities or outcomes

should be equal between people’s lives as a whole or between

3

parts of different people’s lives.

In this paper I address the question of the temporal subject

of justice. My aims are, firstly, to provide a critical

examination of the strongest defences of the lifetime view,

according to which justice requires taking people’s whole lives as

relevant when assessing and establishing their distributive

entitlements and obligations (section 2). Second, I want to

develop an alternative view, on which some time-specific

considerations – that is to say, considerations about how people

fare at specific points in time – have non-derivative weight in

determining what our obligations are to them (section 3). Before

embarking on these tasks, I identify, in section 1, the most

salient considerations that I suggest a defensible view about the

temporal subject of justice should accommodate.

1. Three relevant considerations: responsibility, compensation,

and hardship

That the question of the temporal subject of justice has been

relatively neglected by theorists of distributive justice may be

4

in good part due to the fact that most theorists endorse, most

often without discussing it, a particular answer to that

question, namely, the lifetime view.2 As mentioned above, this is

the view that justice requires taking only people’s whole lives

as relevant when assessing and establishing their distributive

entitlements and obligations. How people fare during parts of

their lives, or at any given moment in time, should not be given

independent, non-derivative weight as a matter of justice; it is

important only derivatively of what ultimately matters, which is

the quality of people’s whole lives. A temporal parts view, by

contrast, holds that what is of fundamental importance is only

1 The first explicit treatment of this question was offered by Norman Daniels in Am I My Parents’ Keeper? (Oxford: Oxford University Press, 1988).I examine the central argument of that book in Section 2 of this paper.2 The lifetime view is endorsed by John Rawls, for whom “the claims of those in each phase [of life] derive from how we would reasonably balance those claims once we viewed ourselves as living through all phases of life,” Justice as Fairness: a Restatement (Cambridge, MA: Harvard University Press, 2001), p. 174. Another proponent is Thomas Nagel, who writes: “the subject of an egalitarian principle is not the distribution of particular rewards to individuals at some time, but the prospective quality of their lives as a whole, from birth to death,” in Equality and Partiality, (New York: Oxford University Press, 1991), p. 69. The lifetime view is assumed also by Ronald Dworkin whenhe writes that his egalitarian “envy test” is to be applied over lifetimes. See Sovereign Virtue (Cambridge, MA: Harvard University Press, 2000), p. 76; Richard Arneson also endorses the lifetime view in “Equality and Equality of Opportunity for Welfare,” Philosophical Studies 6,(1989): 77-93, as does Norman Daniels in Am I My Parents’ Keeper?.

5

how people fare over parts of their lives. On a temporal parts

view, I shall say, there are time-specific obligations of

justice.

It will be helpful to clarify at the outset the nature of

the difference between a lifetime view and a temporal parts view.

The two kinds of view differ over the time-spans over which a

background distributive principle applies, where by “background

distributive principle” I mean an incomplete principle of

distributive justice that specifies only the currency and the

pattern of a just distribution, where the demands of justice thus

conceived are thought to identify people’s fundamental, i.e. non-

instrumental, claims of justice. An example of such a principle

might be the principle that people should, as a matter of

justice, enjoy equality of welfare.3

A temporal parts view maintains that the time-spans over

which a background distributive principle applies are temporal

parts of their lives. It holds that people’s fundamental claims

in distributive justice are to having their conditions over

3 Matthew Clayton and Andrew Williams discuss a number of questions surrounding the currency and pattern of distributive justice in their introduction to their edited collection, The Ideal of Equality (London: Macmillan and St. Martin's Press, 2000).

6

temporal parts of their lives be in line with a particular

pattern of distribution, as measured in a particular currency.

Different temporal parts views identify the relevant temporal

parts of lives in different ways. One such view might hold that

it is simultaneous parts of people’s lives that should be compared

– for example, that the simultaneous parts of different people’s

lives ought to be equal. Another possibility is to hold that it

is non-simultaneous parts that should be compared. An example, here,

would be a corresponding segments view, which holds that we

should compare corresponding parts of different lives – for

example, different people’s childhoods, adulthoods, and old

ages.4 A lifetime view, by contrast, is any view that maintains

that it is people’s entire lives that must be brought into line with

a background distributive principle. So, to illustrate, if one

person has enjoyed more welfare than another during the

simultaneous first halves of their lives, an egalitarian lifetime

view would require that this inequality be reversed during the

second halves of their lives, whereas the simultaneous equality

version of the temporal parts view would not. Note that, apart 4 The “simultaneous equality” and “corresponding segments” views are distinguished by McKerlie in “Equality and Time”.

7

from adopting a temporal parts view on its own (a pure temporal

parts view), or a lifetime view on its own (a pure lifetime view),

we also have the option of endorsing a combined view that

includes both types of view. It is not incoherent, for example,

to maintain of two non-overlapping lives that it is of

fundamental importance that both are equal both at the level of

their whole lives and at the level of their simultaneous parts.

In this section, I suggest that two important considerations

– considerations about responsibility and about compensation –

tell against adopting a pure temporal parts view. But I also

identify a third consideration – which I refer to as the hardship

consideration – which casts doubt on pure lifetime views. This

sets the stage for the rest of the paper, which aims to convert

that doubt into a conviction against pure lifetime views and to

develop an alternative view of the temporal subject of justice.

Before proceeding, I need to state an assumption I will be

making in the arguments that follow. I shall assume that

individuals are personally identical over time, and indeed, in

standard cases, over their entire lives. This assumption is

relevant for a discussion of the temporal subject of justice. To

8

the extent that individuals are not personally identical over

time, the temporal parts view gains in plausibility. For example,

on the view of personal identity defended by Derek Parfit,

according to which personal identity inheres in psychological

continuity over time and according to which such continuity may

not necessarily span the course of a whole life, it may be more

plausible to insist that temporal parts of lives are the temporal

subjects of justice since those temporal parts would map onto

what are, in effect, different persons within those lives.5 While

this suggestion is important, I do not examine it here. The

reasons for this, and for assuming personal identity over time,

are two-fold: first, I do not wish to rule out the lifetime view

at the outset, but to examine its full merits; second, I think

that there are other arguments in favour of the temporal parts

view (at least as a component of a hybrid view) than those that

rest on a relatively controversial picture of personal identity.

It is those arguments I focus on in this paper.6

5 See Derek Parfit, Reasons and Persons, (Oxford: Oxford University Press,1984), Ch. 15. 6 McKerlie also argues that we have personal identity-independent reason to endorse a temporal parts view in “Justice Between the Young and the Old”, pp. 168-174. I discuss McKerlie’s argument in Section 3.

9

(i) Responsibility

It is widely accepted among theorists of justice that a person’s

claims of distributive justice at a given moment in time should

sometimes reflect her earlier exercises of responsibility.

Justice, including egalitarian justice, does not require that we

attend to someone’s disadvantage if she is in the relevant sense

responsible for that disadvantage herself. If someone is badly

off or worse off than others as a result of her own choices, for

example, humanity may tell in favour of helping her, but she has

no claim of justice that others come to her aid. Different

theorists of justice formulate and justify this responsibility

constraint in different ways: some believe it is grounded in a

commitment to the separate principle of desert, others that it is

implied by the demands of equality themselves, properly

understood. Moreover, they identify different necessary and

sufficient conditions for responsibility, and will accordingly

support different judgements as to what disadvantages and

inequalities are compatible with justice.7

These differences, however, should not be relevant here.

10

For any version of the responsibility constraint seems to tell

against our adopting a temporal parts view, or at least, a pure

temporal parts view. According to a temporal parts view, whether

a person’s fundamental claims at a given stage of his life are

satisfied depends only on the nature of his condition during that

stage (or, more fully, on whether his condition during that stage

fulfils a background principle that regulates distribution

between that stage and the equivalent stages of other people’s

lives). A temporal parts view thus “insulates” people’s claims

during those parts of their lives from the impact of their

exercises of responsibility during other parts of their lives.

For example, a temporal parts view that emphasises simultaneous

equality would require that people who are currently better off

should transfer resources to people who are currently worse off,

even if the latter were themselves, during some earlier part of

their lives, responsible for being currently worse off.

7 There is an increasingly large literature on the responsibility consideration that I cannot discuss here. For a representative set of writings invoking the responsibility consideration, see, Richard Arneson, “Equality and Equal Opportunity for Welfare”; G. A. Cohen, “On the Currency of Egalitarian Justice,” Ethics, 99 (1989); Hillel Steiner, “Choice and Circumstance,” in Ideals of Equality, ed. Andrew Mason,(Oxford: Blackwell, 1998); and Ronald Dworkin, Sovereign Virtue (Cambridge: Harvard University Press, 2000), Chs. 2 and 9.

11

Similarly, a corresponding segments equality view that required

equality between people’s childhoods, their middle ages, and

their old ages, would require that an old man be given extra

resources to ensure that his old age is as good as the old ages

of members of the previous generation, even though he may need

those resource transfers as a result of having imprudently

squandered, at some earlier point in his life, the opportunity of

a good old age.

It might be suggested that a temporal parts view can

accommodate responsibility in the following way. The view would

require that there should be simultaneous equality but it would

allow departures from simultaneous equality when these are due to

differential exercises of responsibility in earlier parts of

people’s lives.

We should note two points about this suggestion. First, the

view proposed is not, strictly speaking, a temporal parts view,

since it does not apply a background distributive principle to

temporal parts of people’s lives. A “responsibility-sensitive”

background distributive principle would say that people should be

equal, not in their outcomes, but in their opportunities.

12

Applying that principle to temporal parts would yield the view

that people should have equal opportunities within parts of their

different lives, so that their outcomes within those parts should

reflect their earlier, differential exercises of responsibility

within those same parts. It would not yield the view, proposed in

the previous paragraph, that the outcomes within parts of

different people’s lives should reflect their differential

exercises of responsibility during previous parts of their lives.

But, secondly, although the view does not qualify as a

temporal parts view, we should not exclude it from consideration

merely on that definitional ground. Let us then note, for now,

the possibility of a quasi-temporal parts view that accommodates

responsibility by making distributions between parts of different

lives reflect differential exercises of responsibility during

previous parts of lives. I shall point out below that this view

will still fail to accommodate the third of our guiding

considerations regarding hardship, and that we should look for a

different answer to the question about the temporal subject of

justice.

Considerations of responsibility tell against a pure

13

temporal parts view. On the lifetime view, by contrast, there is

no injustice in simultaneous inequality or in inequality between

key stages of people’s lives, when the people who are worse off

or suffer a disadvantage are themselves responsible for it.8 This

conviction may well account for why people are often drawn to the

lifetime view.

(ii) Moral compensation

A second consideration that informs our judgements about the

temporal subject of justice, and that also tells against a pure

temporal parts view, is that benefits and harms at different

moments in a life morally compensate for each other. A benefit

morally compensates a person for a harm if both, taken together,

satisfy a person’s moral claims (by “moral claims” here I mean

fundamental claims of distributive justice). The possibility of

moral compensation over time is a blessing we often take for

granted. Suppose you told your child that you would spend some

time together this morning, but that you now won’t be able to do

8 As McKerlie puts it, “[t]hinking in terms of lifetimes allows us to take responsibility into account,” in “Justice Between the Young and the Old”, p. 154.

14

that, after all. If moral compensation over time were not

possible, you would not be able to make up for that. But you can

make up for that: you can spend more time with her this afternoon

or this evening or some other time, and her claim to spending a

substantial amount of time with you will have been satisfied that

way.

The moral compensation consideration implies that we ought

to reject pure temporal parts views.9 If benefits and harms at

different times within a life morally compensate for each other,

then whether or not a person’s fundamental claims at a given

stage of his life are satisfied is something that will depend

both on the nature of his condition during that stage and on the

nature of his conditions during other stages of his life.

According to a temporal parts view, however, whether a person’s

fundamental claims at a given stage of his life are satisfied

does not depend on his conditions at any time other than during

that stage. On a corresponding segments view, for example, if a

9 As Thomas Nagel writes, the possibility of moral compensation is a reason “for taking individual human lives, rather than individual experiences, as the units over which any distributive principle shouldoperate,” Mortal Questions (Cambridge: Cambridge University Press, 1979), p. 120.

15

first person’s middle age is less good than a second person’s

middle age, then that should be remedied, even if the first

person’s young age was, or his old age could be, better than the

second person’s young or old age. What’s more, if nothing can be

done to make the first person better off during his middle age,

there would be no reason, on this view, to do anything for him to

offset that fact in his old age, because, on this view, such

moral compensation is not possible. Indeed, the view would

actually imply that if the first person’s old age were made

better off (for no good reason, on the view in question), this

would create a further injustice, for it would make his old age

better than other people’s old ages. The simultaneous equality

view would have similar implications. If, as seems plausible, we

accept that benefits and harms at different times within a life

morally compensate for each other, then we should reject these

pure temporal parts views.10

(iii) The hardship consideration

While considerations about responsibility and compensation push

us away from pure temporal parts views, a question hovers over

16

the proposal that we should endorse a pure lifetime view. The

lifetime view requires that the balance between the benefits

citizens receive across the different stages of their lives be

struck in the way that best secures for them the whole lives to

which it says they have a fundamental claim. But this seems to

entail the following worrying conclusion: it may mean that there

should be no entitlement on the part of citizens to receive

support during a particular phase of hardship so long as the

alternative entitlement they could thereby enjoy would improve

their whole lives more, and even if that alternative entitlement

weren’t necessary to alleviate hardship. I will refer to this

concern as the hardship consideration.

More can be said about the hardship consideration than I

10 The compensation-based argument against the pure temporal parts viewmay need to assume that moral compensation can exist between all phases of a person’s life, since otherwise it does not rule out temporal parts views that identify temporal parts that are shorter than an entire life but long enough to encompass the period within which moral compensation within a life is possible. One might object that moral compensation is not possible between all the phases of a person’s life on the grounds that individuals do not remain personallyidentical over time. As I mentioned earlier, I assume personal identity over time in this paper; the personal identity-based objection to the moral compensation argument therefore does not apply.My reason for assuming this, recall, is that I want to suggest that wehave a personal identity-independent reason for endorsing a temporal parts view (and hence for resisting the moral compensation argument).

17

will here. For my purposes, I need to highlight three main

points. The first is that I shall use “hardship” in a technical

sense, according to which it refers to conditions of persons that we

intuitively believe we have an obligation to help alleviate. Examples of hardship

might include physical or mental suffering and material deprivation, and it

seems increasingly plausible to regard these conditions as

examples of hardship, in my technical sense, the longer their

duration.11 Note, however, that when assessing a given view of

our obligations from the perspective of the hardship

consideration, our focus shouldn’t be on the examples just given,

but on the status they are intended to exemplify – i.e. that of

being of a condition that we intuitively believe we have an

obligation to help alleviate – and it may be that, under certain

circumstances, a person’s suffering or material deprivation, does

not exemplify that status.

The second point is that the hardship consideration I am

levelling against the lifetime view is a version of the challenge

that recent critics of luck egalitarianism like Elizabeth

11 For a valuable discussion of the nature of suffering, see Jamie Mayerfeld, Suffering and Moral Responsibility, (Oxford: Oxford University Press, 1999), Ch. 2.

18

Anderson have levelled against that view.12 There is one

difference, however. Anderson and other critics of luck

egalitarianism seem to think that that view is liable to the

challenge posed by the hardship consideration as a result of

their having constrained their demands of justice by demands of

responsibility. I do not disagree that a theory may become

vulnerable to the hardship challenge as a result of incorporating

responsibility. Indeed, it was precisely on this ground that I

suggested we reject the responsibility-sensitive quasi-temporal

parts view discussed in the previous section. But I am also

suggesting that the hardship challenge is potentially levelled

against any lifetime view, even if we adopt it, not because of

our belief in the importance of responsibility, but because of

our convictions about the possibility of compensation I examined

above.

Thirdly, I would like to emphasise that accommodating the

hardship consideration need not commit us to the implausible view

that citizens must be entitled to support in cases of hardship

regardless of the trade-off such an entitlement would effect at 12 See Elizabeth Anderson, “What’s the Point of Equality?”, Ethics, Vol. 109 (1999): 287-337.

19

other stages of their lives. Still, because hardship is not

something that should be taken lightly, the lifetime view’s

commitment to trading-off entitlements between stages of a life

does place a burden of justification on it. We need to be assured

that the specific way in which the lifetime view commits us to

trading-off entitlements at one stage of life against

entitlements at other stages fits with our intuitions about how

it ought to respond to cases of hardship.

2. The strategy of derivative concern

In this section, I consider a strategy proponents of the lifetime

view might use in order to show that the lifetime view can

accommodate the hardship consideration. They can argue that once

we identify the correct background distributive principle that we

should apply to people’s entire lives, we will notice that it

actually requires, as a means to its fulfilment, that no one is

left to suffer hardship at any stage in their lives. The hardship

consideration can thus be accommodated without abandoning the

view that people’s claims of distributive justice apply, non-

20

derivatively, or most fundamentally, at the level of their entire

lives. People in hardship at any stage of life must always be

supported, not because they have a fundamental claim at each

stage of life to such support, but because this is what will

ensure for them the entire lives to which they do have a

fundamental claim. Let me call this the strategy of derivative concern.

To assess the merits of this strategy, it is helpful to

distinguish between interest-based and ambition-sensitive versions of the

lifetime view. Taking Norman Daniels and Ronald Dworkin as

representatives of these respective versions, I argue that

neither version enables us to prosecute the strategy of

derivative concern.

The interest-based version of the lifetime view

According to Daniels’ prudential lifespan account, people’s

entitlements at different stages of their lives consist of those

resources they would prudently allocate to those stages out of a

“fair lifetime share” of resources, on the assumption that they

are ignorant of their ages and of their conceptions of the

good.13 Daniels characterises “fair lifetime shares” in Rawlsian

21

terms, that is, in terms of expectations of social primary goods

over the course of a whole life. His rationale for appealing to

prudent allocation under conditions of ignorance is simple: such

reasoning will identify a spread of entitlements over people’s

lives that is likely to render their lives better than they might

otherwise be and in a way that is not biased against any specific

age group. His rationale for imposing the hypothetical constraint

of a “fair lifetime share” is that this ensures that people’s

lives are as good as they can be compatibly with treating all

lives fairly.

Daniels maintains that a prudent allocation of our fair

lifetime share over the different stages of our lives is one that

guarantees what he calls age-relative normal opportunity throughout life

– that is, the opportunity to pursue the range of plans that are

normal for a particular stage of life, such as, for example, the

variety of educational paths we might pursue during childhood and

adolescence, or the variety of career or child-rearing plans we

might pursue during adult years. Prudence recommends age-relative

13 For the fullest statement of the account, see Norman Daniels, Am I My Parents’ Keeper?. For a later statement of the account see also, “The Prudential Lifespan Account of Justice across Generations,” in Norman Daniels, Justice and Justification.

22

normal opportunity throughout life because it is possible that we

may revise our plans of life as we age and that we therefore no

longer derive welfare from the plan of life we are, at any given

stage of life, pursuing.14

The general requirement of age-relative normal opportunity

throughout life implies certain concrete entitlements across

different age groups. It implies, for example, that amounts of

health care resources allocated to people in different age groups

may differ substantially, compatibly with the demands of justice,

since maintenance of age-relative opportunity range, to which all

age groups are entitled, may be more expensive during one part of

life than another.

Even if we grant that the specific implications Daniels

draws from the prudential account actually follow from it – that

is, that a prudent allocation of a fair lifetime share over the

different stages of a life would indeed be one that preserves

age-relative normal opportunity range at each stage – the

prudential lifespan account still faces a serious problem.15 It 14 Daniels, Am I My Parents’ Keeper?, p, 76.15 For the objection that the implications Daniels identifies do not follow from his account, see Dennis McKerlie “Equality Between Age-Groups,” pp. 286-8.

23

can yield entitlements that accord with our convictions about

hardship only by restricting the pattern in which our fair

lifetime shares are allocated across the different stages of our

lives in a way that disrespects the autonomy of people whose

ambitions require it to be allocated other patterns. Consider

Hebe, for example, a person who is firmly in the grip of a youth-

orientated conception of the good, so that she prefers, for the

same money-value, less than age-relative normal opportunity while

old in order to have more than age-relative normal opportunity

while young. Assume that Hebe is not asking for something that is

worth more money than her fair lifetime share, so that her having

it would not leave others with less than their fair lifetime

shares. To insist that Hebe’s fair lifetime share take a form

that is insensitive to her ambitions, even when her receiving it

in an ambition-sensitive form would impose no additional costs on

others, seems unjustified.

The ambition-sensitive version of the lifetime view

While an interest-based version of the lifetime view recommends

entitlements that allocate our fair lifetime shares of resources

24

over our lives in a pattern that serves our greatest interest,

the ambition-sensitive version recommends entitlements that

allocate our fair lifetime shares in a pattern that is most

sensitive to our ambitions. The most important example of this

version of the lifetime view is Ronald Dworkin’s theory of

distributive justice, equality of resources.16 The problem with

Dworkin’s theory, I suggest, is that it is incapable of

accommodating the hardship consideration in a robust way.

According to Dworkin, people’s fair entitlements over the

course of their whole lives are “ambition-sensitive” either when

they secure a pattern of resources over their live that reflects

their own choices about how resources should be spread over their

lives and when they compensate people for brute luck in line with

a hypothetical insurance package they would purchase. The

hypothetical insurance cover people are entitled to is worked out

by assuming: (1) people are allocated an equal share of money

with which to purchase insurance to cover their entire lifespans,

(2) they are ignorant of their level of exposure to various risks

against which they might insure, and (3) they purchase insurance 16 Dworkin develops his theory must fully in Sovereign Virtue but see also his responses to critics in “Sovereign Virtue Revisited”, Ethics 113 (2002)

25

in line with their own ambitions.17 We should, for example,

determine the kinds of publicly provided health care to which

people are entitled, and the tax contributions they must make

towards it, by determining the kind of health insurance package,

and corresponding premium, they would select for their whole

lifetime under those three hypothetical conditions.18

Dworkin’s ambition-sensitive version of the lifetime view

clearly upholds the responsibility and compensation

considerations. It requires that people’s claims at any point be

sensitive to their own exercises of responsibility by allowing

their outcomes at any given moment in time to reflect either

their earlier choices, or their own preferences for insurance.

Furthermore, it upholds the compensation consideration by

insisting that people make their choices and are covered by 17 Dworkin uses the hypothetical insurance model in order to determine compensation for a number of different kinds of brute luck, including disability, unemployment due to lack of talent, and poor health. See Sovereign Virtue, Chs. 2, 7 and 9. The differences between these various kinds of brute luck and insurance are not relevant for my discussion here. For helpful discussion of Dworkin’s insurance model see Colin Macleod, Liberalism, Justice and Markets (Oxford: Oxford University Press, 1998).18 A key reason for why an ambition-sensitive lifetime view should employ the hypothetical insurance model is that our ambitions in life are profoundly informed by our attitudes to risk, and that our entitlements thus reflect our ambitions fully only when they also reflect our attitudes to risk. See Sovereign Virtue, Chs. 7 and 9.

26

insurance on the basis of equal lifetime shares of money.

Consider Hebe again, who wants to dedicate a substantial amount

of her equal share of resources to the earlier parts of her life,

in full knowledge that this will mean that she will have only

very few resources available while old. While this decision may

seem imprudent, it should be allowed on Dworkin’s view.

The problem with Dworkin’s version of the lifetime view is

that it seems incapable of meeting the hardship consideration.

That consideration, recall, is that there are certain conditions

of persons that we intuitively believe we are obligated to help

alleviate. Now, as we have seen, according to Dworkin, the kinds

of resources that must be available for people who suffer through

bad luck is something that must be determined according to the

hypothetical insurance model. And while it is likely that many

people would like to avoid many forms of conditions that

constitute hardship in my technical sense, it is by no means

clear that they will want to avoid all of them. It is quite

possible, for example, that some people’s ambitions would not

lead them to provide enough insurance for their very old age, say

in the form of adequate long-term care, and yet, despite that

27

fact, we would still believe, I submit, that we are obligated to

alleviate the conditions they would have to endure without such

care. The ambition-sensitive approach therefore fails to

adequately accommodate the hardship consideration.19

Dworkin might reply to this objection in several ways but

none are convincing. First, he can emphasise that since basing

welfare provision on each individual’s hypothetical insurance

choices would involve prohibitive information-gathering costs,

people’s welfare claims are to be modelled in a statistical way,

that is, on “what people of ordinary opinion and prudence would

have done in the hypothetical circumstances.”20 That approach, so

it might be suggested, would yield conclusions that fit better

with our intuitions about the obligations we have to alleviate

hardship. 19 Many commentators on Dworkin have identified as problematic his theory’s inability to compensate for some forms of bad brute luck we may find problematic. For this line of criticism, see G. A. Cohen, “Onthe Currency of Egalitarian Justice,” Ethics 99 (1989): 906-944, Michael Otsuka, “Luck, Insurance, and Equality,” Ethics 113 (2002): 40-54, and Robert van der Veen, “Equality of Talent Resources: Procedures or Outcomes,” in Ethics, 113 (2002): 55-81. While these critiques of Dworkin focus on the theory’s putatively insufficient egalitarian credentials (in that it does not fully neutralise the unequal impact of bad brute luck) my focus, like Anderson’s (see her “What is the Point of Equality?”) is on the fact that it cannot even securely guarantee help for certain absolute harms.20 See Dworkin, “Sovereign Virtue Revisited”.

28

This reply does not escape the objection raised two

paragraphs ago. To point out that the insurance approach in its

statistical form insures against a large part of those conditions

that we intuitively believe we ought to help alleviate for each

other, is not to vindicate that approach. It may just mean that

it converges in large part with some other standard for

identifying those obligations. What needs to be further shown is

that in cases in which the insurance approach does not match up

with our intuitions, it is our intuitions, and not the approach,

that we would be prepared to revise. That, however, does not seem

to be the case: even if people on average would not insure

against long-term care, we would still, I submit, believe

ourselves to be obliged to alleviate the conditions the elderly

would have to endure without such care.

Second, Dworkin makes two further points that might be seen

as accommodating the hardship consideration. He says, first, that

even if people’s welfare claims were modelled on their insurance

choices, paternalistic considerations would justify not asking

them to internalise the costs of their choices when these would

lead to hardship; and secondly, that a concern with protecting

29

the interests of others who would be called upon to respond to

their hardship would justify forcing everyone to insure so that

each provides for his own relief from hardship.21

We should note the following concerns about these two

points. First, as far as the paternalism Dworkin invokes is

concerned, it seems desirable for a view to avoid reliance on it

if possible. Although Dworkin does not find the paternalism

involved here problematic, it would be preferable if we could

robustly accommodate the hardship consideration without relying

on premises that alternative views of the value of freedom would

reject, and it is an open question as to whether there is such a

possibility.

Second, as far as Dworkin justifies a system of compulsory

insurance that avoids hardship out of a concern for the interests

of those who would otherwise be called upon to support those in

hardship, we need an account of why others would indeed have any

such obligation towards those in hardship.

In section 3 below I put forward a view that answers both of

these points at once. It gives a non-paternalistic justification 21 Dworkin makes this point in his reply to Elizabeth Anderson’s critique. See “Sovereign Virtue Revisited”, p. 114.

30

for why we should prevent hardship, which seems to me to capture

the urgent moral concern that makes us respond to hardship in the

way we do. Moreover, the view I shall offer also accounts for why

others would indeed have to respond to hardship, so that a

concern with their interests can be said to underwrite compulsory

insurance. My view, however, is a departure from a pure lifetime

view, which, as I have shown in this section, cannot by itself

accommodate the hardship consideration, at least not without

resorting to paternalism.

Having considered the extent to which two leading versions

of the lifetime view enable us to derive appropriate concern for

people facing hardship, we can now see that the failure of that

strategy is not specific to these two versions, but applies to

all versions of the lifetime view. The strategy of derivative

concern must (a) provide an attractive account of the lifetime we

must help secure for each other and (b) derive from it the

conclusion that we must support each other during hardship at any

stage of our lives. The problem with the strategy is that it

cannot do both of these things. Dworkin’s ambition-sensitive

version of the lifetime view is attractive insofar as it

31

describes that lifetime as one that is lived in the light of our

own ambitions. But embracing this account comes at the expense of

our being unable to guarantee support for persons in hardship. On

the other hand, while Daniels’ interest-based version of the

lifetime view may succeed in providing that guarantee, it can

achieve that result only by describing the lifetime we must

secure for each other in a way that is partially insensitive to

the ambitions we may wish to pursue in life. Stated most

generally, then, the reason the strategy of derivative concern

fails for all lifetime views is this. No lifetime view can both

insist that the lifetimes we must help to secure for each other

are lifetimes that reflect our own, different ambitions in life,

and that, as a means to fulfilling that end, we must guarantee

support for each other during the hardship we might experience at

any stage of our lives. Embracing the pure lifetime view must

therefore come at the expense of the hardship consideration.

3. Dignity and time-specific constraints

In this section I argue that we should combine and constrain the

32

lifetime view with a temporal parts view that is grounded in an

account of dignity. We must assume, as the lifetime view demands,

that each person is to be allocated a fair lifetime share of

resources but we must insist that each person’s dignity

constrains how people’s fair lifetime shares of resources may be

allocated over the different stages of their lives. More

specifically, I will argue that dignity

constrains us to ensure that our fair lifetime shares are

allocated over our lives in such a way that either (a) we can

always endorse our life plans or (b) we always have a decent set

of opportunities at each stage of our lives. Combining the

lifetime view with a dignity-based constraint provides, I

believe, a more promising way of accommodating our three guiding

considerations about responsibility, compensation and hardship

than the competing views I examined above.

The account of dignity I will sketch is not intended to

square with all of the many ways in which we use the concept of

dignity in ordinary speech or in philosophical treatments of the

subject. The term “dignity” has been used to refer to a status

that turns on a person’s rank in a social hierarchy, or on the

33

extent to which a person conducts herself in a morally upright

way, or on the degree to which a person displays fortitude in the

face of adversity, or simply on whether she or he is a human

being.22 The sense in which I shall use “dignity” covers much,

but not all, ordinary usage, and follows a broadly Kantian

tradition, in which the term refers to a person’s capacity for

living autonomously.23 My aim is not to offer a comprehensive

justification for the interpretation of dignity I shall be

relying on, but only to show that there exists at least a

plausible account of dignity, which, when combined with a

lifetime view, would enable us to satisfactorily accommodate the

three guiding considerations about how we should answer the

question about time.24

The account of dignity proposed here consists of three

central claims. The first two have been put forwarded by many

others, and although they are not uncontroversial, I shall not

attempt to add to the arguments that have been made in their

support. The third claim in my account is a distinctive one (so

far as I am aware) and I shall therefore devote more time to

providing support for it.

34

(i) Intrinsic value

Dignity is a name for the intrinsic value we, as persons, have,

in virtue of our (internal) capacity to direct our lives in the

light of our appreciation of value. The notion of intrinsic value

is complex, but I invoke it here to convey only this point: our

capacity for autonomy is “non-prudentially valuable” – that is,

it is valuable independently of whether and how it is valuable

for the person whose dignity is in question. Some might deny this

and insist that our dignity is valuable only because it is ether

22 Lennart Nordenfelt distinguishes four meanings of dignity in “The Varieties of Dignity”, Health Care Analysis, Vol. 12, (2004): 69-81; For a historical overview of the concept of dignity see Herbert Spiegelberg,“Human Dignity: A Challenge to Contemporary Philosophy,” in R. Goteskyand E. Lazlo (eds.) Human Dignity: This Century and the Next (London: Gorden and Breach, 1970); for a conceptual analysis, see Aurel Kolnai, “Dignity”, Philosophy, Vol. 51, (1976): 251-271. For the connection between dignity and fortitude, see David Beyleveld and Roger Brownsword, Human Dignity in Bioethics and Biolaw, (Oxford: Oxford University Press, 2004), p. 253. 23 For Kant’s view that autonomy is the ground for dignity see his Foundations of the Metaphysics of Morals (Upper Saddle River, NJ: Prentice-Hall,1997), p. 53. For more recent discussions of the relationship between autonomy and dignity, see Thomas E. Hill Jr., Autonomy and Self-Respect (Cambridge: Cambridge University Press, 1991) and Robin S. Dillon (ed.), Dignity, Character, and Self-Respect (London: Routledge: 1995), pp. 14-18.24 Although I shall not discuss the specific lifetime view with which we should combine the dignity-based temporal parts view I shall be proposing, the rationale for the latter view will tell in favour of certain lifetime views over others, namely those that accommodate autonomy. I explain the connection between dignity and autonomy below.

35

instrumentally necessary for, or constitutive of, our welfare. My

account of dignity rejects that claim and assumes that dignity is

valuable even when it does not stand to our welfare in either of

these ways, although of course it may also do that.25 In the next

sub-section, I explain why my temporal parts view must insist

that dignity is an intrinsic value if it is to succeed in

accommodating our three guiding considerations regarding

responsibility, compensation and hardship.

(ii) The internal capacity for autonomy

The ground of a person’s dignity – the property that gives her

dignity – is her internal capacity for autonomy, by which I mean

the various psychological conditions that are necessary for her

to be able to direct herself in the light of her appreciation of

value, such as, for example, her capacity to critically reflect

upon the ambitions she wishes to pursue in life, and her capacity

to revise those ambitions in the light of new evidence. Three

points need to be emphasised here. First, how exactly we should

25 For the view that autonomy is valuable only as a constituent elementof welfare, see John Stuart Mill, On Liberty, ed. Elizabeth Rapaport (Indianapolis: Hackett, 1978), Ch. III.

36

characterise the internal capacity for autonomy has been the

subject of prolonged controversy.26 My account of dignity will

remain agnostic with respect to that controversy. Secondly, by

the “capacity for autonomy” we might also mean to refer to

certain conditions external to the person, such as freedom from

interference by others or the availability of a range of

meaningful options for action. My account excludes such external

conditions from the ground of a person’s dignity, in the sense

that whether or not someone has the ground of dignity does not

depend on, and does not vary with whether, she has access to

certain external conditions for realising her autonomy. Were

these to be included, this would yield the perverse implication

that a person who has the internal capacity for autonomy but who

is interfered with, or deprived of meaningful options, lacks the

properties that confer dignity upon her. In denying that freedom

from interference and the availability of meaningful options are

part of the ground of a person’s dignity, we of course do not

imply that those external conditions cannot be part of what a

26 For an overview of controversy, see John Christman, “Constructing the Inner Citadel: Recent Work on the Concept of Autonomy”, Ethics 99 (1988): 109-124.

37

person’s dignity demands that others secure for her. Indeed, as I

shall argue below, respecting people’s dignity does require,

typically, that we secure for people certain external conditions

at all stages of their lives.

A third point addresses the concern that an autonomy-grounded

account of dignity is too narrow in a couple of respects. It might be

suggested, firstly, that children and persons with serious cognitive

impairment still possess dignity even if they are not fully capable of

autonomy. Secondly, it might be suggested that we can violate a

person’s dignity without undermining her capacity for autonomy, for

example by ridiculing her in a way that involves treating her as a mere

object. Both suggestions imply that the ground of dignity is different

from, or at least broader than, the capacity for autonomy.

In response to this concern notice, first, that objections can be

raised against both suggestions just made about the narrowness of

autonomy as a ground for dignity. One might insist, firstly, that our

disinclination to refuse children and the cognitively impaired the

status of dignity discloses not our belief that they have dignity, but

that others hold strong moral obligations towards them, a belief that

is compatible with maintaining that children and the cognitively

38

impaired have a moral status different to that of dignity. Against the

second suggestion, one could argue that to ground dignity in autonomy

in no way implies that the duties dignity demands are restricted to

duties of non-interference with autonomy but allows that those duties

include the duty to refrain from symbolically denigrating people in

treating them as if they did not possess autonomy, as we would do, for

instance, if we treated them as mere objects.

A second response to the concern about narrowness is less

controversial. Even if it were true that the capacity for autonomy is

too narrow a ground for dignity, this would not undermine the proposal

developed here that dignity generates a plausible constraint on the

lifetime view. For that proposal need presuppose only that what

dignity demands includes respect for autonomy, in a sense I outline

immediately below, and not also that its ground is restricted to

respect for autonomy.27

(iii) Time-specific autonomy

The first two claims in the account of dignity sketched so far –

27 I would like to thank an anonymous referee for pressing me to address the concern that autonomy may be too narrow a ground for dignity.

39

namely, that dignity is an intrinsic value of persons and that it

is grounded in the capacity for autonomy – follow a broadly

Kantian tradition of thinking. Where my account is unusual is in

its third claim, which further specifies the nature of the

capacity for autonomy in which dignity is grounded and teases out

the temporal dimension of that capacity. So far, I have referred

to that capacity as the capacity to give shape to one’s life in

the light of one’s own appreciation of value. Now, so described,

that capacity is still open to interpretation: it might be

interpreted, first, as the capacity of a person at one particular stage

of her life to give shape to the life of that person at all other

stages (for example, the capacity of a young adult to give shape

to the life that lies ahead). Alternatively, it might be

interpreted as the capacity of a person at each and every stage of her

life to contribute to giving shape to her life as a whole. To

illustrate, consider a case in which a person, at an earlier

stage of her life, contemplates pursuing a life plan that may

severely restrict her opportunities at a later stage of her life.

If her dignity inheres in the capacity for autonomy interpreted

in the first of our two senses – i.e. as a capacity of an earlier

40

self to give shape to the life of that person overall would not

disrespect her dignity. But if her dignity inheres in the

capacity for autonomy interpreted in our second sense – i.e. as a

capacity of a person at each and every stage to contribute to

giving shape to a life – it may.

It is a crucial element of the temporal parts view defended

here that we should understand a person’s dignity in the second

sense, that is, as inhering in a person’s capacity for autonomy

at each moment in her life. What might be said in favour of this

time-specific interpretation of the capacity for autonomy?

Consider the following example. Imagine a society in which

everyone, near the beginning of their adult lives, makes a set of

choices about how their whole lives should unfold. This society

has set aside a week for this purpose – Lifetime Decision Week – during

which all 21-year olds must decide the kind of career they would

like to pursue during their lifetimes, whether they want to

marry, whether they want to have children, what kind of religion

they will follow, if any, and so on. Suppose now that at the end

of the week, once all young adults have made their choices about

these things, they voluntarily take a pill that ensures that they

41

never change their minds about their choices in the future. In

this society, people’s lives will be, in good measure, the

results of their own choices.28 But it is difficult to accept

that their lives embody the value of autonomy, and indeed, it is

more plausible to say that they have acted in a way that offends

against that value. This intuition, I believe, illustrates the

appeal of time-specific autonomy: we do not lead autonomous lives

by having our earlier selves make choices that affect and bind

our later selves, independently of the attitude which our later

selves have towards those choices, so that our whole life, or as

much of it as possible, is the result of an early choice. Rather,

what matters, for our autonomy, is that our lives are endorsed as

they are lived, and that we continuously affirm our plans of life

as we pass through all the stages of our lives.28 One might object that what is disturbing about the idea of a lifetime decision week is only the fact that it results in important decisions being made in a hurried way, because they must all be taken together at once, rather than their being the result of proper deliberation, and in response to whatever opportunities arise organically, as one moves through life. The reply to this objection isthat lifetime decision week would still be disturbing even we imagine that individuals make very thoughtful and highly sensible decisions about their lives on the basis of extremely detailed information abouttheir future prospects. Lifetime decision week is disturbing not merely because it might lead to imprudent decisions, but also because it prevents individuals in the future from re-assessing their lives inthe light of the values they will come to appreciate.

42

Which constraints does the value of dignity impose on how

people’s fair lifetime shares must be allocated over the

different stages of their lives? I will not attempt to answer

this question comprehensively but only to the extent necessary

for showing that the value of dignity, when combined with a

lifetime view, gives us the most plausible answer to the question

about justice and time.

A person’s dignity is protected at a given time when she can

live in the light of her appreciation of value, and hence

autonomously, at that moment in time. For this ideal to be

achieved, it is necessary that one or the other of the two

following conditions is met, where each condition, on its own, is

sufficient for dignity.

First, a person’s dignity is respected if that person

endorses the plan of life that has placed her in the

circumstances she finds herself in at a given moment in time. By

“endorsing” a plan of life, I mean “deciding in favour of it

given the alternative life plans she might have pursued”. A

person who endorses a life plan regards a difficult condition she

might find herself in at a given moment in time as a “price worth

43

paying” for that life plan, where her assessment of whether it is

a price worth paying takes place against a background assumption

of how many resources are available to her overall.29

Furthermore, to endorse a life plan does not imply that one is

actively pursuing it. It may be the case that a person is no

longer pursuing a life plan, e.g. that of having a satisfying

career, perhaps because she no longer can, or because she no

longer values that sort of thing, but she can still endorse that

plan in the sense that she can recognise that it played a

valuable role in her life given the type of person she was when

she pursued it.

The claim that endorsement of one’s life plan at a given moment in

time is sufficient for dignity needs to be clarified in the following

way. What is being claimed as sufficient for dignity is a person’s

endorsement of a life plan that allocates a fair lifetime share of resources over

her life . It is not sufficient for her dignity that she endorse just any

29 Recall that we are considering here, what role dignity-based considerations play in setting time-specific constraints on a fair lifetimeshare view. So it is legitimate, on this combined view, to hold that genuine endorsement depends on a person’s having had available a fair lifetime share of resources on the basis of which she might have pursued alternative life plans, and furthermore, that it does not depend on her having had available to her more than her fair lifetime share.

44

life plan. It is possible that a person might endorse a life plan that

places her in a servile, or, in a some other way inferior, position to

someone else, e.g. a life plan that involves always serving the

interests of her husband at the expense of her own, no matter what the

relative importance is of their respective interests. That a person

endorses such a life plan is not sufficient for her dignity. My claim

that endorsement is sufficient presumes that a person is selecting

between life plans that allocate a fair lifetime share of resources

over her life, a presumption that does not hold in the case of the

servile wife.30

Secondly, a person’s dignity is protected if that person has

access to a decent set of opportunities to engage in valuable

activities at each particular stage of life, regardless of

whether the first condition above is satisfied, that is,

regardless of whether that person endorses, at any particular

stage of her life, the life plans that have given shape to her

life overall and affected that particular stage in the way they

have.31 Like the endorsement condition, the decent opportunities

30 I would like to thank an anonymous referee for pressing me to clarify the endorsement condition. For a related discussion of servility and its relationship to dignity see Thomas Hill, Autonomy and Self-Respect (Cambridge: Cambridge University Press, 1991), Ch. 1.

45

condition is sufficient for dignity, and it is justified on the

basis of the underlying idea that a person’s dignity inheres in

her capacity at each moment in time to live in the light of her

appreciation of value. If a person has decent opportunities at a

given moment in time, she need not endorse the life plan that has

affected her present condition in order for her dignity to be

protected. To illustrate, consider a person who no longer

endorses her plan of life – if she could go back in time, she

would chose a different plan. Nevertheless, she has plenty of

opportunities available to her at this stage of her life. Imagine

she regrets having saved up to such an extent that she now, in

the final decade of her life, lives in luxury, when she might

have used up some of those savings, with better results, earlier

on in her life. Were we to assume that the endorsement condition

is a necessary condition for dignity, we would have to conclude,

bizarrely, that her dignity is not protected in this last decade

of her life. We should instead say that her ample opportunities

are sufficient for her dignity to be protected.

Accommodating the three guiding considerations

46

I believe that a combination of the dignity-based temporal parts

view and a lifetime view constitutes the most plausible basis for

accommodating our three guiding considerations about justice and

time. Let us first consider the hardship consideration. The

dignity-based temporal parts view constrains us to ensure that

our fair lifetime shares are allocated over our lives in such a

way that either (a) we can always endorse our life plans or (b)

we always have a decent set of opportunities at each stage of our

lives. Now, it is impossible to demonstrate in the abstract that

this test ensures that our fair lifetime shares are allocated in

such a way as to protect us against all cases of hardship, but it

is difficult, I submit, to think of troubling cases it would

allow. In particular, it seems fair to say that physical or mental

suffering or deep material deprivation would be ruled out by the dignity-

based test. Most people would not regard these conditions as

prices worth paying for their life plans, and these conditions

would also severely limit their opportunities. The dignity-based

31 While I do not think there is a precise formula for guiding our judgements about when a person’s opportunities are so meagre as to threaten her dignity, I think we can form reasonable judgments about when a person’s set of opportunities are so small that we can say “herlife is no longer unfolding in light of what she deems valuable”.

47

temporal parts view would thus constrain us to allocating our

fair lifetime shares over our lives in a way that prevents these

conditions from occurring.

It might be objected, however, that there are some cases of

hardship that would not be ruled out by my account of dignity, in

the sense that the latter will fail to support obligations to

people at least in some cases of hardship. This objection might

be pressed in two ways. First, my account of dignity allows

people to remain in conditions that result from life plans they

endorse, and it seems possible that people can endorse life plans

that leave them in conditions we would classify as hardship, such

as severe pain or poverty, for example. Here, my response is to

bite the bullet: if a person genuinely endorses her life plan

(and recall the above requirement that she must have had a fair

lifetime share available with which to pursue alternatives) –

that is, if she would genuinely have pursued the same life plan

again – then, while helping to alleviate her pain or her poverty

might be good thing to do, it is not, I submit, a moral

obligation, and her condition thus does not amount to a case of

hardship in my technical sense – that is, as a condition we

48

intuitively believe we have a moral obligation to alleviate.

The other way the objection might be pressed is by

suggesting that it is possible that a person could be in severe

pain or deep poverty while having a decent set of opportunities

available to her, in which case, again, my account would say that

we have no obligation towards her (on grounds of dignity). My

response here is to deny the compatibility of conditions we would

classify as hardship, such as severe pain or deep poverty, with a

person’s having a decent set of opportunities available to her.

It is of course possible for people to soldier on with a backache

in their successful careers and family life, but if they can do

this, then the pain is not severe enough to count as hardship.

Let me next consider how including the dignity-based

temporal parts view as part of our answer to the question about

the temporal subject of justice enables us to accommodate the

responsibility and compensation considerations, both of which,

recall, militate against temporal parts views in general. Here

the claim that dignity is of intrinsic value plays a crucial

role.

Notice, to begin with, that the conflict between the

49

responsibility and compensation considerations, on the one hand,

and our obligation to respect people’s dignity, on the other, is

not a necessary one, in the sense that it is not bound to occur

in all possible scenarios. In fact, it is possible to arrange

things so that that conflict disappears, or at least so that the

scope of cases in which it exists diminishes significantly. More

specifically, it is possible to remove options from people, which

options, if people exercised them, would threaten their future

autonomy. For example, one can prohibit people from gambling to

the extent of leaving themselves destitute, or one can prohibit

people from spending income earlier in their lives that they

could have saved up for their pensions. Removing such options

would not go against the responsibility and compensation

considerations: these two considerations do not speak to the

range of options that should be available to people, but require

only that certain consequences should follow in case certain

options are exercised. If the relevant options are removed from

people, we can then honour the responsibility and compensation

considerations without leaving people in conditions in which they

live without dignity.

50

The question, of course, is whether it is indeed legitimate

to remove options from people in the ways just suggested. It is

here that the fact that dignity is an intrinsic value plays a

crucial role. For suppose, for a moment, that dignity mattered

only because of the impact respecting it has on our welfare. If

that were the case, it would be difficult to justify removing

options from people on the grounds that these would threaten

future dignity. A person might level the same complaint against

this proposal that he could level against the interest-based

version of the lifetime view we examined above in section 2: it

is either paternalistic or otherwise disrespectful of autonomy to

insist that he refrain from acting, or from receiving portions of

his fair lifetime share in ways that harm only his own welfare.

However, if, as I am maintaining, a person’s dignity has

intrinsic value, then removing options that threaten his dignity

can be justified on grounds that are welfare-independent and

respectful of autonomy. Let me explain the nature of that

justification more fully.

Consider, first, a simple way in which one might seek to

appeal to the intrinsic value of dignity in order to justify

51

removing options from people. One might argue that we may remove

options from people that threaten their dignity simply because we

may prevent people from responding inappropriately to matters of

intrinsic value. In other words, the argument is that we may

force people to maintain a proper relationship to matters of

intrinsic value. Now that argument faces a powerful objection. It

is not obvious why someone’s failing, or the prospect of his

failing, to respond appropriately to an intrinsic value calls for

forcible prevention, as opposed merely to moral condemnation, at

least in cases in which that failure does not harm others.

Indeed, one might claim that, in such cases, it ought to be a

matter for each person’s conscience how he interprets, respects

or seeks to realise an intrinsic value in his life. Consider, for

example, the intrinsic value that is so central to the abortion

controversy, the sanctity of human life. As Dworkin argues, it

may be more plausible to maintain that people on opposite sides

of that controversy ought to tolerate each other in their

different, conscientious interpretations of human sanctity, than

to maintain that one side may impose, through the apparatus of

the state, its own interpretation of that value on the other

52

side.32

Instead of appealing to the claim that we may force people

to realise certain intrinsic values in their lives, we can appeal

to a second and more plausible argument for why matters of

intrinsic value, such as respect for dignity, justify removing

options from people. A person who puts himself in a situation in

which he lives without dignity confronts others with a dilemma.

They can either ignore his condition, in which case they fail to

abide by a moral obligation they have to respect his dignity, or

they can abide by that obligation at the cost of having to

relinquish some of their resources to him. In light of this, it

may be reasonable for others to remove options from a person so

as to protect themselves from being confronted with that

dilemma.33 Notice that this argument does not assume, as did the

simpler argument of the previous paragraph, that the intrinsic

value of a person’s dignity just by itself justifies the forceable

removal options from that person. It assumes only that that

32 I am grateful to Andrew Williams for alerting me to the relevance ofDworkin’s argument in this context. See Dworkin, Life’s Dominion (London: Harper Collins, 1993), p. 101. 33 I have developed this argument more fully in Paul Bou-Habib, “Compulsory Insurance without Paternalism” Utilitas 18 (2006): 243-263.

53

intrinsic value generates a non-enforceable moral obligation of

support on others, and it gives us the conclusion that they may

forceably remove options from a person, by adding the additional

assumption that they may limit their exposure to incurring that

moral obligation of support.

One might object that this argument for why the intrinsic

value of dignity allows us to prevent people from leaving

themselves in conditions of hardship is incongruous with the

underlying spirit of the hardship consideration: the latter

points to the fact that we must be concerned for the conditions

of others, whereas the argument I have just sketched focuses on

the interests of those who witness their conditions. When we

learn of people in hardship, so the objection might go, our moral

attention is captured, surely, by their hardship, and the threat

to their dignity it involves, not by the burden that alleviating

their hardship would impose on us. Notice, however, that nothing

in the argument I have just put forward disagrees with that.

Indeed, the argument crucially depends on the claim that we have

a strong moral obligation to help others who cannot live with

dignity, and it is not incompatible with that claim to maintain

54

that we may reasonably prevent others from acting in ways by

which they will foreseeably impose that moral obligation on us.

Dignity vs. time-specific priority

Before concluding, I would like to strengthen the appeal of my

proposal by arguing that it is more plausible than the only other

combined view that has been proposed. According to McKerlie, we

should combine what he calls the “lifetime priority view” with

the “time-specific priority view”.34 Both views are versions of

the more general view known as prioritarianism. According to

prioritarianism, the moral value of assigning a benefit is

greater, the worse off the condition of the person receiving the

benefit (worse off, that is, in absolute terms). Now, McKerlie

believes this principle applies both when the “worse off

condition” refers to the condition of a person’s whole lifetime,

and when it refers to the condition of a part of a person’s life.

So, he believes that prioritarianism needs to be applied twice,

to lifetimes and to parts of lives, and that we must find some

way of weighing lifetime priority against time-specific priority

when they conflict.34 See Dennis McKerlie, “Priority and Time”.

55

Combining a lifetime view (whether lifetime priority or any

other) with the time-specific priority view is a promising way of

accommodating the consideration about hardship. If the moral

value of assigning a benefit is greater, the worse off the part

of life to which it is assigned, the conclusion follows that

greater benefits must flow to those who face hardship at any

time. However, I believe that the time-specific priority view

provides a less plausible basis for accommodating the

responsibility consideration than a dignity-based temporal parts

view.

In the previous section, I argued that a dignity-based

temporal parts view can plausibly be reconciled with the

responsibility consideration. We can ensure that dignity is

respected compatibly with holding people responsible by removing

options from people. Furthermore, because our removing those

options is justified by appeal to an intrinsic value grounded in

our autonomy, we would not incur the charge of failing to respect

autonomy. It is difficult to see how the time-specific priority

view can be reconciled with the responsibility consideration in

the same way. Time-specific priority demands a response to

56

persons who have low levels of welfare at specific times. In view

of this, it cannot justify our removing options from people on

welfare-independent grounds. Indeed, to uphold time-specific

priority, McKerlie’s view may be liable to the very same kind of

objection he has levelled against Daniels’ prudential lifespan

account.35

4. Conclusion

From a certain point of view, it may appear irrational for us to

establish entitlements for age groups in a way that fails to

ensure that our whole lives go as well as possible. Such policies

may appear to rest on little more than an emotional response to

the vivid hardship we see people having to endure in the here and

now. We see residents in care homes having to endure a life of

very limited means, or victims of Alzheimer’s gradually

enshrouded in the darkness of their disease. It is difficult

35 It is open to McKerlie to argue that (1) time-specific priority neednot be welfarist but could be autonomy-focused, so that it requires prioritising support for those most lacking in autonomy at a specific time, and (2) that autonomy is intrinsically valuable. These two points would bring his combined view closer to mine.

57

under such circumstances not to endorse generous welfare

policies. It may be said that we should resist this tendency, and

broaden our focus by endorsing entitlements in such cases only if

they form part of an optimal lifetime’s set of entitlements we

should design for all citizens.

In this paper I have argued that the fact that our moral

attention is captured by the hardship people undergo during parts

of their lives may in fact be the proper response to the

intrinsic value of their dignity. That point has not been

sufficiently acknowledged, I believe, by proponents of the

lifetime view, and it leaves them incapable of plausibly

accommodating our convictions about the appropriate response we

should have to the hardship of others. That it is intrinsically

important that a person’s dignity always be respected does not

mean, of course, that the state should refrain from designing

entitlements in terms of their impact on the quality of people’s

whole lives. But it does mean that the latter is not the only

concern in view of which those entitlements should be designed.36

36 I would like to thank the following people for their comments on previous drafts of this paper: Richard Arneson, Fabian Freyenhagen, Dennis McKerlie, Serena Olsaretti, Andrea Sangiovanni, Andrew Williamsand two anonymous referees.

58

59